Noisy Laptop Power Supply.

Go To Last Post
36 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

My 12-year old Laptop power supply was replaced several years ago with an unbranded 19.5V / 4.74A from Ebay with 5.5mm/2.5mm barrel jack.

The PC works fine with this PSU.

 

Now that I have my smart Rigol DS1054Z I see horrible interference which disappears if I run the Laptop off its battery.

 

If I buy a used HP 19.5V/4.74A will this cure my problem?

 

Or can anyone recommend a suitable PSU?   Preferably new.

 

Or should I just buy a new Laptop?    (this one was Vista-32.  currently Win7-32)

 

David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

david.prentice wrote:
Now that I have my smart Rigol DS1054Z I see horrible interference which disappears if I run the Laptop off its battery.
... and worse from there (wall wart over-volts, excessive ripple causes the device to anomalously reboot)

Wall warts can be a bane.

david.prentice wrote:
If I buy a used HP 19.5V/4.74A will this cure my problem?
Maybe

Have a 4.5 year old HP notebook PC wall wart that's still working well.

david.prentice wrote:
Or can anyone recommend a suitable PSU?
Not me

At your preferred electronics distributor, browse the wall wart datasheets for

  • warranty's duration
  • ripple
  • switching frequency (optional, likely not stated)
  • lightning suppression (or consider the MOV in the outlet strip)

david.prentice wrote:
Or should I just buy a new Laptop?
Yes though this 12yo notebook PC will be an excellent tool to operate when FUBAR occurs.

Some USB Type-C PD have over-voltage protection (OVP) and appears to have a power switch capability that's in only a relatively few USB Type-A.

Sometimes one wants to power cycle the MCU or MPU that's on USB VBUS.

Sometimes one requires more than 5V (USB Type-C PD is max 20V 5A though most wall warts are less)

I dis-like USB Type-C as its retaining force decreases with use (I cycle the connector once or twice per day, the battery indicator shows when I inadvertently dis-connect by a bit of motion)

Am considering USB Type-C to Type-A high retention (an industrial USB connector, increased force, increased temperature range)

 


C-Type USB - USBGear

High Retention USB Connectors - Kycon | Mouser

 

edit : wall warts can have an instability

https://www.avrfreaks.net/forum/xmega-sram-slow-turnaround-solved-glitchy-power-supply#comment-1028546

 

edit2 : USB Type-C PD is common in Chromebooks and most recent Windows notebook PC as this eases docking.

https://www.avrfreaks.net/forum/chrome-os-third-os#comment-2756086

...

... both devices will offer LTE options and USB-C docking.

...

 

"Dare to be naïve." - Buckminster Fuller

Last Edited: Sun. Sep 1, 2019 - 03:20 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You could try plugging into a line filter, though the issue could be radiated emissions (like a radio transmitter).

 

If you move the laptop/supply further away, while leaving it plugged in as-in....does the interference reduce?

 

 

I have my smart Rigol DS1054Z I see horrible interference 

If your scope leads are just floating (disconnected from anything), that itself is not a good test...of course, a good scope will sniff out any signals in the ether.

If you see noise during a circuit measurement, then that is a problem.

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:
If you see noise during a circuit measurement, then that is a problem.
A common note in wall wart datasheets ripple data is the scope's bandwidth is limited to 20MHz with two specified capacitors on the load (for USB Type-A that's 10uF max then typically a 100nF, ripple measured at a significant load current)

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I see 500mV pk-pk noise on a proper circuit measurement.  i.e. Arduino Zero powered by Laptop USB.   10X probe with GND clip on Zero GND.

 

If I run the Laptop on battery the noise disappears.   The battery lasts for about 20 minutes.  (Zero takes about 20mA.  USB Mouse takes about 0mA)

 

If I power the Laptop from a different Mains receptacle (different room) it is still noisy.

If I unplug the 19.5V/4.7A from the Laptop I get "less" noise.  e.g. random 100mV pk-pk.

 

When the Laptop battery is "reporting short life" the noise is horrendous when I plug the barrel-jack in to recharge.  2000mV pk-pk.

 

David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

david.prentice wrote:
I see 500mV pk-pk noise on a proper circuit measurement.

...

2000mV pk-pk.

Ouch

One wall wart spec's 50mVpp for 5Vdc 1.6A

 

This Raspberry Pi 4 wall wart (5.1V 3A) is nearly 5 times the ripple with OVP at 10V :

T7715DV TT Electronics - IoT Solutions | Mouser

 

A USB Type-C PD wall wart that has better performance :

SWC45-P59121520-NB CUI | Mouser

 

A USB Type-C PD sink controller has OVP at 6 V for 5V (+/-20% for any USB Vbus of 5V, 9V, 12V, 15V, 19V, 20V)

EZ-PD™ Barrel Connector Replacement (BCR) (Cypress Semiconductor)

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0



 Arduino Zero powered by Laptop USB

Well, you left out an EXTREMELY important point---that the laptop is powering your circuit, not simply being in the vicinity. Indeed, it could very well be the switcher noise of the laptop brick.

To be more accurate, this is the external adapter/charger, not the power supply of the laptop itself (internal).

 

What if you run the laptop on batt & plug in the charger (NOT into the laptop, leave it on battery) and put it close by the scope setup?  Does the noise return.

Noise can  also differ due to the fact the scope is grounded, but the laptop may not be, since it may completely float (batttery power), or semi float (via the brick), or be well-grounded (via the brick).   Some brick's neg output connects to earth gnd & some don't

 

If you could run the rigol off of battieries, you could tell whether it was a charger noise vs gnd issue (most likely noise).

 

You could add some filters to the charger output.

 

 

 

 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

david.prentice wrote:

If I unplug the 19.5V/4.7A from the Laptop I get "less" noise.  e.g. random 100mV pk-pk.

 

Yes,   the 19.5V/4.7A brick is powered.  The barrel jack unplugged.   So the Laptop is ruunning off the battery.

 

If I run from battery for 20 minutes and then plug in the barrel jack,   I get the horrendous 2000mV noise.    It takes 20 minutes to recharge the battery to 100%.

The noise is still bad when the battery is 100%.

 

I think that I will just buy the used HP branded 19.5V/4.7A brick.    It comes from

We are a Computer Refurbishing company based in the WestMidlands.  All items listed are fullytested and refurbished where required by our in-house technicians.

 

I can't believe that any genuine brick could be as noisy as my current one.

 

If I plug the Zero into the Rigol USB I get a clean trace.

The Laptop is sitting next to the Rigol.   Both Rigol and Laptop brick are powered from the same mains outlet.

 

David.

Last Edited: Sun. Sep 1, 2019 - 07:48 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have a 90W HP AC Adaptor Model PA-1900-08H1 (19V 4.74A)

If I measure between barrel jack GND (GND because my meter probe won't fit down the socket hole) and Mains earth (AC range non RMS meter)  I measure 1.97VAC.

 

Make of that what you will.

 

PS. Different mains sockets give different results and testing the +ve part of the barrel gives 0.9VAC. Interestingly different DMM range gives either 0V or 11V. My take on that is that there is definitely some noise from the HP adaptor but it is too high a frequency for my DMM to present a proper measurement.

 

Last Edited: Sun. Sep 1, 2019 - 09:11 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The scope is grounded, and the so is the computer when the wall wart is connected, these connect the ground in a loop. You need to break that loop, I wish I knew how to explain the magic, but it would be rambling BS about magnetic fields and stuff I don't understand.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

and the so is the computer when the wall wart is connected

Well maybe, maybe not

 

Some chargers plug into the wall with a two prong cord, and some with a three prong (grounded) cord.  Even then, it's not a certainty!   

Take a hammer & open up the charger to check, just to be safe.

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think it is probably fair to say that there are a large number of replacement no-brand laptop chargers out there which signally do *not* behave as the EC/ULA markings printed on them suggest they ought to. In particular, radiated noise in he MHz to 3GHz band can be horrendous - the actual limits are quite low (shades of the recent certification thread?).

 

A switchmode power supply module running with no load is often noiser than one with a load, due to the random stop-start operation - but lots of cheap power supplies radiate like mad: both electromagnetically and on the input and output lines... You need to pay attention to your power supply because it's a prime location for noise to be injected.

 

Neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have just borrowed a 19V / 3.16A brick from my next door neighbour.   It has a 5.5mm/2.5mm barrel jack.

 

I plugged it in to my (charged) Laptop.   Zero noise on the scope.

I will run the Laptop on battery for 20 minutes.   Then see if the (smaller) PSU can recharge it ok.

 

The 19V / 4.74A brick has a clover leaf mains lead.  i.e. it uses the mains GND pin.

The 19V / 3.16A brick has a two-pin mains lead.  i.e. with plastic GND pin.

 

I can't remember when I bought the unbranded 19V / 4.74A brick.

It always powered the Laptop ok.   And with several Arduinos on the USB ports.

But I had never looked at scope signals.

 

I am fairly confident that the "smaller" PSU can keep the Laptop battery charged.

The real test is with a sustained recharge of a flat battery.

 

David.

 

Edit.

The Laptop ran on its battery for about 20 minutes i.e. 100% -> 7% -> LOW BATTERY WARNING

After 20 minutes of recharging it reports 7% -> 20%.    The brick is warm to the touch.  The scope is noise-free.

After 60 minutes of recharging it reports 7% -> 30%.    The brick is warm to the touch.  The scope is noise-free.

I will shut the lid down.   That should give the PSU an easier job.

 

I suspect that it will take a lot of minutes to recharge to a full 100%.

The battery icon is on for 1 sec.  off for 5 sec.   Probably means the PSU is worried about temperature.

 

I think that it is worth buying a branded 4.74A brick.    It will probably be both noise-free and capable of safely recharging (the knackered battery).  

Last Edited: Mon. Sep 2, 2019 - 09:11 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm surprised the noisy 4.74A brick charges the battery to 100% in 20 minutes, but the 3.16A brick only gets it to 30% in an hour.  Somehow that just sounds wrong to me.  What am I missing?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Let's put some guesswork numbers in.   e.g. 2A for PC operation.  2A to recharge battery.

 

If the PC is hibernating,  the 3.16A brick supplies 0A to PC and can recharge the battery with up to 3A continuously.

If the PC is active,  the 3.16A brick overheats.   It supplies current for 1 second  (probably far in excess of 3.16A).   The battery supplies current for 5 seconds.

 

The Laptop is designed for a 4.74A brick.    It can supply current to the PC and current to the battery.

 

No.   I have not measured the current from either brick.    I am just guessing that the PC Battery firmware is asking for more current than the 3.16A brick can provide without overheating.    Hence 1 sec on / 5 sec off duty cycle.

 

My initial theory was that a 100% charged battery would mean the 3.16A brick only needs to trickle charge the battery.   And could supply active PC + trickle charge without getting warm.

In practice it does get fairly warm.    Unwise to run the Laptop from the 3.16A brick.

 

But at least I am convinced that a brand-name brick is quieter than my Ebay unbranded brick.

When the 4.7A HP brick arrives,   I will be happy.     (and will dismantle my noisy brick to see how it is made)

 

David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I dis-like USB Type-C as its retaining force decreases with use 

USB cables with jack screws (Type-A, Type-B, Type-C)

USB Cable - National Instruments

...

The USB A plug features a jackscrew that securely screws into the USB retention mount next to each USB port on supported CompactRIO and stand-alone CompactDAQ systems and meets requirements for Class I, Div/Zone 2 hazardous locations.

...

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

From my experience with switched mode power supplies, its usually the capacitors that die. It comes down to the quality of the caps used and the design. Some designs are more conservative than others.

 

For the 'average' brick psu - these generally have a UC3842 psu controller chip running a mosfet in a flyback converter arrangement. To start the controller up, there is a resistive dropper to provide enough current to get things going, then there's a winding on the transformer, diode and a 10uF (or thereabouts) cap that takes over when the converter is up and running. The resistive dropper was a common cause of failure as it is one or more high value resistors. Due to the voltage imposed across it (320VDC for 240VAC systems), you get electromigration in the resistor and it fails. Most designs are aware of this and use a string of resistors in series to negate this problem thus spreading the voltage across a number of resistors. The other common problem is the 10uF cap dries out - this causes the supply to 'tick'. Then there's the input and output filter caps which eventually die. They normally get a 'roof top' due to the pressure build up. If this is happening, then they are about to die. That sums up around 90% of the cause of failure. 

 

 

Other things with the brick supplies that have caused issues for me in the past are:

1. Grounded/non grounded

2. if grounded, is the output 0V tied to ground (my collegue found this out the hard way recently - he blasted the USB port of his new laptop!)

3. if not grounded, there's a Y cap from the 'live' side to the 0V side to shunt the transformer capacitance from coupling the switching noise. This causes a small leakage current that can upset things and give you a little tingle. My macbook pro has this. If you have a part back to real ground, this causes current to flow and can upset ADC readings etc if you're not careful. It can also zap your electronics.

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:

2. if grounded, is the output 0V tied to ground (my collegue found this out the hard way recently - he blasted the USB port of his new laptop!)

On my example of the HP adaptor (probably identical to OP's failed unit) I measured the resistance between mains earth on the clover-leaf socket and barrel jack outer and discovered approx 5.6 ohms. There's almost certainly a current in this resistor which may explain the voltage measureents in my post #9.

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Dictionary of power plugs:

 

1. two pin - 'figure 8'

2. three pin clover leaf - 'mickey mouse'

3. usual IEC - 'jug plug'

 

Just so we're clear.

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You can also get very strange results when your unit is plugged in "over here" and your scope is plugged in "over there".  Ideally they should be plugged into the exact same outlet or power strip (and not one in the outlet & the other in the power strip that's plugged into the outlet).   

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

 

In the U.S., we refer to the three pins on the AC plug as Hot, Neutral, and Ground. Sometimes the Ground is more accurately called a "Safety Ground".

On the DC Plug going into the laptop, the ground conductor is usually (but not always) connected to the Safety Ground. Doing this makes it easier to pass regulatory testing approval (EMC, ESD, Etc.). It also reduces leakage current through the "isolation" barrier inside the power supply. Some people complain that when laptops are placed on their bare laps, they get a tingling sensation where the screws of the laptop touch their skin. Having those screws connected to the safety ground of the AC plug mostly gets rid of that issue.

Also, even if the safety ground is not connected to the ground conductor of the DC jack, it could be providing some additional safety benefits inside of the power supply itself.

 

https://electronics.stackexchange.com/questions/78079/why-is-this-laptop-adapter-grounded

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Tue. Sep 3, 2019 - 08:17 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:
Ideally they should be plugged into the exact same outlet or power strip (and not one in the outlet & the other in the power strip that's plugged into the outlet).   

I had that configuration with both the scope and the laptop brick plugged into the same power strip, and it took me a while to figure out why my scope was showing the same large amount of noise on everything I measured.  When I unplugged the brick everything was back to normal.  I dont remember what I did with that brick.  I think I put it in the recycle box.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well, so much for conducted emissions standards!

 

Unlike CE, US has no conducted or radiated emissions susceptibility standards. The standard FCC disclaimer is that if you have problems with some device that is sensitive  to external signals, YOU have to fix it because the "emitter" is within standards. But, even CE susceptibility targets functionality: does the scope stop responding to changes in front panel knobs in the presence of line noise, etc. The CE standards, as I recall them, are really borderline on things like noise on a video display so I would expect that a scope showing noise on its display would be pretty much ho-hum (both literally and figuratively).

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The scope is a 100 MHz Tektronix 2235.  Is that ho-hum?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hardly, but for the standards people, it probably is.

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I wasn't sure what you were saying.  I thought you were bad mouthing my scope, which I rather like.  It has no fan noise, unlike the digital Tektronix I have and almost never use, and a giant, ancient Tektronix with the small green display, which I haven't used in a long time, but it fits in the roll around stand that came with it, and i have the other two scopes strapped to the top of it.  I have to admit I am sort of a snob when it comes to scopes and greatly prefer the analog to the digital.  Sure, digital scopes can do lots of stuff an analog scope cant and are required at times, but that digital layer between the in and the out bothers me.  For every day use I like my signals undiluted.  I like vinyl records too.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

No, I really like Tek analog scopes. I worked for then during that era. 

 

I was trying to poke at the viewpoint of the  standards setters. Tek scope? Ho-hum, its just another susceptible appliance!

 

One of the  things that MIGHT be happening is common-mode noise. Tek analog scopes were pretty well filtered at the power entry, but common mode was a bit weak. Think of line and neutral moving together relative to safety ground (but, oops, the UK folks don't have that arrangement). Then, when you connect the signal input (e.g. probe) to a circuit that has an AC path to ground, you can get AC currents in the probe ground and those currents generate voltages that appear as spurious signals. 

 

One of the things that you CAN do is to build a filter box. This might work for David, also. Think of a little outlet box with a short power cord. Inside the box, add a common mode choke and a low-pass filter that has its primary corner at a few KHz and is good to 100MHz or so (not totally trivial). Bet it would reduce the displayed noise, substantially.

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I will just wait for the used HP 4.7A brick to arrive.
Then I will open the noisy brick.
I suspect that it either omits all filter components (as shown in some Youtube videos)
Or it has components mounted on the pcb but failures as described by Russell in #17.
.
Meanwhile I will just avoid the Laptop.
.
At least I know that my neighbour's HP 3.16A brick is noise-free. So brand name bricks should be ok.
.
David.

Last Edited: Tue. Sep 3, 2019 - 09:52 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I haven't had any problems with display noise besides that one experience.  I was pretty confused that the noise I was seeing was everywhere on everything I measured.  When I unplugged the laptop brick all was peaceful again.  I figured it was getting into the scope ground and moving everything around.  Since then all has been copacetic.  I really like that scope.  You did a good job on it.  When it dies I will probably get another one just like on eBay.  They are only a couple hundred dollars

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0



When it dies I will probably get another one just like on eBay. 

 I had almost 40 of these Fluke digital scopes, which I still occasionally use (though I generally prefer color). The real nice thing is these are also analog scopes....switching modes  goes from "digital  noise"  to an analog trace that is as clean as a whistle.  not bad for 4 channels.  Sometimes these go for $50, though typicall a few hhundred$$.

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I suspect that those Fluke scopes were rebranded from some other source. But, not sure. But they are not selling that line any more. All of their "scopes" are sort of super-multimeter size now.

 

Re: Tek 2235 - had nothing to do with that line. That was after I left. I worked on 580/585 (summer intern stuff) then a couple of years later, the 5000-series (low-cost plug-in series parallel to 7000 series, due to many complaints about high purchase cost of 7000's).

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think Philips rebranded their test gear to Fluke in the later years. The Philips stuff wasn't too bad for its time and it was lower end professional, but we all know the ultimate was Tek. A friend advised 'get the best you can afford' - I got a second hand 2465. Never regretted it. Still works, although I've had to change the odd cap in the power supply and a CA3046 that went weird.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0



I suspect that those Fluke scopes were rebranded from some other source

They bought/merged with Philips in the 80's/90's (the "famous" test alliance).  Fluke also bought/rebranded Gigatronics RF generators.  Ah, the good old days 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

david.prentice wrote:
At least I know that my neighbour's HP 3.16A brick is noise-free. So brand name bricks should be ok. . David.

 

Brand name supplies generally have to work, and meet standards, world-wide. That does tend to improve the breed.

 

neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The used HP 4.74A brick arrived in the mail.    Unfortunately it had a 5.0mm / 2.0mm plug.  
 

The used HP brick has a shielded cable with 5.0mm / 2.0mm plug.

The unbranded brick had an unshielded 2-wire cable with 5.5mm / 2.5mm plug.

I cut off the 5.0mm and 5.5mm from new and old cables.    And joined the 5.5mm plug to the HP cable.

 

The unbranded unshielded cable would be one source of noise.

I suspect that there is "another" problem i.e. visible 2000mV noise is not good (tm)

 

I opened the case.    Nothing obvious had blown.  

I have hacked the plastic case,   broken glued items,  hacked the cable.    I will just bin the "remains".

 

Oh.   The HP 4.74A is showing no noise on the Rigol.

There is one happy bunny in Wormshill.

 

I will discharge the battery and see how it performs with recharge one day.

Since the battery is knackered I am unlikely to ever use on battery power.

 

David.

Last Edited: Wed. Sep 4, 2019 - 02:34 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The reason may be the ground loop you create with the scope's GND/PE connection and the USB GND - even if the laptop supply is not GND connected there are capacitors which make a HF connection to the PE of the mains plug or to the life wires of the PSU. Other PSU's may be not so noisy, but a better solution should be a proper (short and thick) connection of the USB GND with the scope's GND (NOT only via the 10:1 probe's coaxial cable). With a isolation transformer for the scope you would be able to prevent the ground loop.

have a nice day